Matthias Boehm created SYSTEMML-2182:
----------------------------------------
Summary: Univariate stats over 10Kx1K w/ wrong optimizer choice
Key: SYSTEMML-2182
URL: https://issues.apache.org/jira/browse/SYSTEMML-2182
Project: SystemML
Issue Type: Sub-task
Reporter: Matthias Boehm
The parfor optimizer decided incorrectly on remote parfor for this very small
input dataset.
{code}
Total elapsed time: 38.600 sec.
Total compilation time: 1.978 sec.
Total execution time: 36.622 sec.
Number of compiled Spark inst: 1.
Number of executed Spark inst: 1.
Cache hits (Mem, WB, FS, HDFS): 4/0/0/426.
Cache writes (WB, FS, HDFS): 2/0/2.
Cache times (ACQr/m, RLS, EXP): 3.799/0.001/0.006/0.341 sec.
HOP DAGs recompiled (PRED, SB): 0/0.
HOP DAGs recompile time: 0.000 sec.
Spark ctx create time (lazy): 21.545 sec.
Spark trans counts (par,bc,col):0/0/0.
Spark trans times (par,bc,col): 0.000/0.000/0.000 secs.
ParFor loops optimized: 1.
ParFor optimize time: 0.076 sec.
ParFor initialize time: 0.000 sec.
ParFor result merge time: 8.191 sec.
ParFor total update in-place: 0/0/0
Total JIT compile time: 26.803 sec.
Total JVM GC count: 8.
Total JVM GC time: 0.87 sec.
Heavy hitter instructions:
# Instruction Time(s) Count
1 ParFor-ESP 26.999 1
2 uacmax 0.762 1
3 > 0.184 1
4 write 0.102 1
5 rand 0.006 1
6 - 0.001 4
7 * 0.001 5
8 mvvar 0.000 23
9 uamax 0.000 1
10 rmvar 0.000 6
{code}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
